Journal of Plant Nutrition, 33:1696–1711, 2010
C Taylor & Francis Group, LLC
Copyright ISSN: 0190-4167 print / 1532-4087 online
DOI: 10.1080/01904167.2010.496892
NITROGEN USE EFFICIENCY IN UPLAND RICE GENOTYPES
N. K. Fageria, O. P. de Morais, and A. B. dos Santos
National Rice and Bean Research Center of EMBRAPA, Santo Antônio de Goiás, Brazil
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
2
Nitrogen (N) deficiency is one of the most yield-limiting nutrients in upland rice growing regions
word wide. A greenhouse experiment was conducted with the objective to evaluate nineteen upland
rice (Oryza sativa. L.) genotypes for N use efficiency. The soil used in the experiment was an Oxisol
and two N levels used were without N application (low level) and an application of 400 mg N kg −1
of soil (high level). Grain yield and yield components and N uptake parameters were significantly
affected by N and genotype treatments. Regression analysis showed that plant height, shoot dry
weight, number of panicles per pot, number of grains per panicle, grain harvest index, N uptake in
shoot and grain were having significant positive relation with grain yield. Nitrogen concentration
of 6.4 g kg −1 in the shoot is established as deficient level and 9.5 g kg −1 as sufficient level at
harvest. Agronomic efficiency of N (grain yield/unit of N applied) and N utilization efficiency
(physiological efficiency X apparent recovery efficiency) were significantly different among genotypes.
These two N use efficiencies were having significant quadratic relationship with grain yield. Soil
pH, exchangeable soil Ca and base saturation were having significantly positive association with
grain yield. However, soil extractable phosphorus (P), potassium (K), hydrogen (H +), aluminum
(Al) and cation exchange capacity were having significantly negative association with grain yield.
Keywords: agronomic efficiency, grain yield, physiological efficiency, yield components
INTRODUCTION
Rice is a staple food for more than 50% of the world population. A major
part of rice production is produced and consumed in Asia (Fageria, 2001a).
Asia, China, and India are major rice-producing and -consuming countries.
There are two main ecosystems of rice, known as upland and lowland. Classification of rice ecosystems is based on environmental conditions. The main
environmental conditions considered in classifying rice ecosystems are soil
moisture regime, soil drainage, land topography and temperature (IRRI,
1984). Upland rice also known as aerobic rice is mainly grown in Latin
Received 27 November 2008; accepted 26 July 2009.
Address correspondence to N. K. Fageria, National Rice and Bean Research Center of EMBRAPA,
Caixa Postal 179, Santo Antônio de Goiás, GO, CEP 75375-000, Brazil. E-mail: [email protected]
1696
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Nitrogen vs Upland Rice Genotypes
1697
America, Africa and Asia. Upland rice is direct seeded in non-flooded, well
drained soil on level to steeply sloping fields.
Upland rice yield is much lower than lowland-flooded rice. The main factors which are responsible for the lower upland rice yield are water deficit and
use of low inputs by the farmers. These inputs include fertilizers, fungicides,
insecticides, and herbicides. Use of low inputs is associated with drought risk
and economic considerations. In spite of low yield, upland rice will continue
to play an important role in rice growing regions due to lack of irrigation
facilities and low production cost. Among essential plant nutrients, nitrogen
(N) is one of the most yield limiting nutrients for upland rice production.
The N deficiency in upland rice is related to low organic matter content
of rice growing soils, soil acidity, soil erosion, use of low level of N fertilizers by farmers due to high cost of these fertilizers. Nitrogen deficiency
is also related to low N use efficiency by the crop due to loss by leaching, volatilization, disnitrification and erosion (Fageria and Baligar, 2005).
Hence, use of N efficient genotypes in conjugation with use of chemical fertilizers is an important complementary strategy in improving rice yield and
reducing cost of production. Upland rice genotypes differ significantly in N
uptake and utilization efficiency (Fageria et al., 1995b; Fageria, 2007). The
objective of this study was to evaluate upland rice genotypes for N use efficiency and determine relationship between yield and yield attributing plant
parameters.
MATERIALS AND METHODS
A greenhouse experiment was conducted at the National Rice and Bean
Research Center of EMBRAPA, Santo Antônio de Goı́as, Goı́as, Brazil to evaluate N-use efficiency of 19 upland rice (Oryza sativa L) genotypes. The soil
used in the experiment was an Oxisol (Typic Haplustox). It had the following chemical and textural properties before the application of N treatments:
pH 5.1 (1:2.5 soil-water ratio), extractable phosphorus (P) 0.8 mg kg−1,
extractable potassium (K) 86 mg kg−1, extractable calcium (Ca) 1.4 cmolc
kg−1, extractable magnesium (Mg) 0.6 cmolc kg−1, extractable aluminum
(Al) 0.2 cmolc kg−1, extractable copper (Cu) 3.6 mg kg−1, extractable zinc
(Zn) 1.2 mg kg−1, extractable iron (Fe) 76 mg kg−1, extractable manganese
(Mn) 45 mg kg−1, and organic matter 20 g kg−1 of soil. Textural analysis values were 469 g kg−1 clay, 180 g kg−1 silt and 351 g kg−1 sand. Phosphorus and
K were extracted by the Mehlich 1 extracting solution [0.05 M hydrochloric
acid (HCl) + 0.0125 M sulfuric acid (H2 SO4 )]. Phosphorus was determined
colorimetrically, and K by flame photometry. Calcium, Mg, and Al were extracted with 1 M potassium chloride (KCl). Aluminum was determined by
titration with sodium hydroxide (NaOH), and Ca and Mg by titration with
ethylenediaminetetraacetic acid (EDTA). Micronutrients were determined
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
1698
N. K. Fageria et al.
on a portion of the extract of P by atomic absorption spectrophotometry.
Organic matter was determined by the Walkley-Black method, and textural analysis by pipette method. Soil analysis methods used in this study are
described in a soil analysis manual published by EMBRAPA (1997).
The treatment consisted of two N levels, i.e., without N application (low)
and an application of 400 mg N kg−1 (high) of soil using urea and 19 upland
rice genotypes. The high level of N was applied in two equal split applications. One-half at sowing and the remaining half at active tillering (43 days
after sowing). Genotypes used in the experiment were: CRO 97505, CNAs
8993, CNAs 8812, CNAs 8938, CNAs 8960, CNAs 8989, CNAs 8824, CNAs
8957, CRO 97422, CNAs 8817, CNAs 8934, CNAs 9852, CNAs 8950, CNA
8540, CNA 8711, CNA 8170, BRS Primavera, BRS Canastra, and BRS Carisma.
These are the promising genotypes supplied by the breeders of the National
Rice and Bean Research Center. A complete randomized design was used
in a factorial arrangement, and treatments were replicated three times. The
study was conducted in plastic pots, with 5 kg of soil in each. At the time of
sowing, in addition to N treatments, each pot received a basal application
of 200 mg P kg−1 as triple superphosphate and 200 mg K kg−1 as potassium
chloride (KCl). These fertilizer rates were based on the recommendations of
Fageria and Baligar (1997b). Each pot contained four plants, and soil moisture was maintained at about field capacity during the experimentation.
At maturity, number of panicles, dry matter of shoot, grain yield, number
of grains per panicle, weight of 1000 grain, spikelet sterility, grain harvest
index and N harvest index was determined. Dry plant material (shoot and
grain) was dried in a forced-draft oven at about 70◦ C until of a constant
weight, and milled. Total N in the plant tissue was determined with a Tecator
1016 digester and 1004 distilling unit according to method of Bremner
and Mulvaney (1982). Grain harvest index and nitrogen harvest index were
calculated by using the following formulas (Fageria et al., 1997):
Grain harvest index = Grian yield/Grain and shoot yield
Nitrogen harvest index = Nitrogen uptake in grain/Nitrogen uptake in grain
and shoot
Definitions and N equations for calculating N use efficiencies are given
in Table 1.
All data were analyzed by analysis of variance, and the F-test was used to
determine treatment significance. Turkey’s test was used to compare treatment means at 5% probability level. Appropriate regression equations were
also used to further analysis of relations between grain yield and N uptake
parameters and soil properties.
Nitrogen vs Upland Rice Genotypes
1699
TABLE 1 Definitions and methods of calculating nutrient use efficiency
Nutrient efficiency
Agronomic efficiency (AE)
Physiological efficiency (PE)
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Agrophysiological efficiency
(APE)
Apparent recovery efficiency
(ARE)
Utilization efficiency (EU)
Definitions and formulas for calculation
The agronomic efficiency is defined as the economic production
obtained per unit of nutrient applied. It can be calculated by:
AE (mg mg−1) = Gf –Gu /Na , where Gf is the grain yield of the
fertilized pot (mg), Gu is the grain yield of the unfertilized pot
(mg), and Na is the quantity of nutrient applied (mg)
Physiological efficiency is defined as the biological yield obtained per
unit of nutrient uptake. It can be calculated by:
PE (mg mg−1) = BYf –BYu /Nf –Nu, where, BYf is the biological yield
(grain plus straw) of the fertilized pot (mg), BYu is the biological
yield of the unfertilized plot (mg), Nf is the nutrient uptake
(grain plus straw) of the fertilized pot, and Nu is the nutrient
uptake (grain plus straw) of the unfertilized pot (mg)
Agrophysiological efficiency is defined as the economic production
(grain yield in case of annual crops) obtained per unit of nutrient
uptake. It can be calculated by:
APE (mg mg−1) = Gf –Gu /Nuf –Nuu , where, Gf is the grain yield of
fertilized pot (mg), Gu is the grain yield of the unfertilized pot
(mg), Nuf is the nutrient uptake (grain plus straw) of the fertilized
pot (mg), Nuu is the nutrient uptake (grain plus straw) of
unfertilized pot (mg)
Apparent recovery efficiency is defined as the quantity of nutrient
uptake per unit of nutrient applied. It can be calculated by:
ARE (%) = (Nf –Nu /Na ) × 100, where, Nf is the nutrient uptake
(grain plus straw) of the fertilized plot (mg), Nu is the nutrient
uptake (grain plus straw) of the unfertilized pot (mg), and Na is
the quantity of nutrient applied (mg).
Nutrient utilization efficiency is the product of physiological and
apparent recovery efficiency. It can be calculated by:
EU (mg mg−1) = PE × ARE
RESULTS AND DISCUSSION
Plant height was significantly (P < 0.01) influenced by N and genotype
treatments and across two N levels varied from 93 to 118 cm among genotypes
with an average value of 103 cm (Table 2). There was a significant quadratic
relationship between plant height (X) and grain yield (Y = −196.3034 +
3.2542X – 0.0097X2, R2 = 0.65∗∗ ). Hence, grain yield increases with increasing plant height but there is a limit of this increase. When plant height is
too short it produces less dry matter and when too high it may lodge and
less responsive to N fertilization (Yoshida, 1981). This means intermediate
plant height is a better compromise and this is confirmed by the highest
grain yield producing genotype CNAs 8993 (at high N level) that did not
have the highest plant height but this genotype had an intermediate plant
height. Shoot dry weight was not affected significantly among genotypes at
low N level. However, Shoot dry weight of 19 genotypes was significantly (P <
0.01) influenced by high N level and varied from 44.4 to 75.8 g per pot at the
1700
N. K. Fageria et al.
TABLE 2 Plant height across two N levels and shoot dry weight and grain yield at two N levels
Shoot dry weight
(g pot−1)
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Genotype
CRO 97505
CNAs 8993
CNAs 8812
CNAs 8938
CNAs 8960
CNAs 8989
CNAs 8824
CNAs 8957
CRO 97422
CNAs 8817
CNAs 8934
CNAs 9852
CNAs 8950
CNA 8540
CNA 8711
CNA 8170
BRS Primavera
BRS Canastra
BRS Carisma
Average
F-Test
N Level (N)
Genotype (G)
N×G
CV (%)
Grain yield (g pot−1)
Plant height (cm)
across two
N levels
Low N
level
High N
level
Low N
level
High N
level
106cde
104cde
93g
99efg
112abc
99efg
94g
105cde
108bcd
109bcd
105cde
104cde
106cde
93g
118a
93g
115ab
94fg
102def
103
15.7a
14.9a
17.1a
13.9a
14.9a
14.5a
14.1a
18.7a
16.4a
15.2a
16.0a
13.9a
15.3a
14.9a
16.9a
17.3a
14.5a
16.2a
14.1a
15.5
53.8defg
52.3defg
62.8bcd
52.2defg
46.1defg
49.7efg
47.0fg
44.4g
49.9efg
56.0cdef
59.0bcde
44.9fg
48.6efg
58.9bcde
54.2defg
75.8a
49.0efg
67.6ab
66.7abc
54.7
14.2a
16.1a
12.6a
12.0a
11.9a
13.7a
12.9a
12.6a
10.8a
15.9a
10.3a
8.4a
11.4a
10.5a
12.8a
12.1a
14.1a
9.1a
8.9a
12.1
61.2ab
69.6a
60.9ab
55.6ab
55.3ab
63.2ab
46.9ab
57.3ab
50.5ab
51.7ab
50.9ab
54.2ab
53.4ab
54.8ab
51.5ab
37.7c
59.3ab
52.5ab
57.2ab
54.9
∗∗
∗∗
∗∗
∗∗
∗∗
∗∗
∗∗
7
12
NS
4
∗∗
Means followed by the same letter in the same column are not significantly different at the 5%
probability level by Turkey’s test.
∗ ∗ , NS
Significant at the 1% probability level and nonsignificant, respectively.
high N level (Table 1). At high N level, shoot dry weight was 253% higher
compared to low N level across genotypes. Shoot dry weight (X) showed
highly significant quadratic relationship with grain yield (Y = −27.7437 +
3.0163X – 0.0269X2, R2 = 0.92∗∗ ). Yoshida (1981) reported that rice yield
can be increased with increasing dry matter yield but relationship is not linear. Similarly, Kiuchi et al. (1966) reported that the increase of grain yield
is certainly brought about by the increase in total dry matter yield, however
at the same time it is found that the increment rate of the grain yield has a
gradual decreasing tendency with the increase in the total dry matter yield.
Fageria and Baligar (2001) also reported a quadratic relationship between
rice grain yield and dry matter yield of shoot.
Grain yield was significantly (P < 0.01) influenced by N and genotypes
treatments as well as by N X genotype interaction (Table 1). At low N
level grain yield did not differ significantly among upland rice genotypes.
1701
Nitrogen vs Upland Rice Genotypes
TABLE 3 Panicle number, panicle length and grain harvest index of nineteen lowland rice genotypes
at two N levels
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Panicle no. pot−1
Panicle length (cm)
Grain harvest index
Genotype
Low N
level
High N
level
Low N
level
High N
level
Low N
level
High N
level
CRO 97505
CNAs 8993
CNAs 8812
CNAs 8938
CNAs 8960
CNAs 8989
CNAs 8824
CNAs 8957
CRO 97422
CNAs 8817
CNAs 8934
CNAs 9852
CNAs 8950
CNA 8540
CNA 8711
CNA 8170
BRS Primavera
BRS Canastra
BRS Carisma
Average
8.0ab
8.3ab
10.3a
7.7abc
6.0bc
8.0ab
9.3ab
6.7abc
7.3abc
8.0ab
8.7ab
3.7c
6.7abc
8.7ab
8.7ab
10.0ab
7.7abc
8.0ab
8.3ab
7.9
17.7fg
22.0a–f
28.0ab
20.3c–g
15.0fg
21.7a–g
26.3abcd
17.0fg
18.7efg
19.3defg
21.3b–g
14.3g
19.7c–g
25.3a–e
21.0b–g
29.0a
17.0fg
25.3a–e
27.0abc
21.4
17.9bcd
16.7de
16.2def
17.5cde
20.8a
16.3def
17.6cde
19.6abc
16.8de
20.5ab
17.6cde
18.6abcd
17.3cde
13.9f
17.4cde
18.8abcd
19.7abc
17.6cde
15.3ef
17.7
23.8abcd
19.5efgh
19.1fgh
20.6defgh
26.6a
18.9fgh
19.2efgh
25.4abc
23.9abcd
25.3abc
21.6c–h
22.6b–g
23.1a–e
18.4h
22.8a–f
21.2d–h
25.7ab
20.9d–h
18.7gh
22.0
0.48a
0.52a
0.43a
0.43a
0.45a
0.49a
0.47a
0.40a
0.40a
0.51a
0.39a
0.36a
0.43a
0.41a
0.42a
0.41a
0.49a
0.36a
0.37a
0.43
0.54abcd
0.57a
0.49abcd
0.52abcd
0.54abc
0.56abc
0.50abcd
0.56ab
0.50abcd
0.48abcd
0.46cd
0.55abc
0.52abcd
0.48abcd
0.49abcd
0.33e
0.55abc
0.44d
0.47bcd
0.50
F-Test
N Level (N)
Genotype (G)
N×G
CV (%)
∗∗
∗∗
∗
∗∗
∗∗
∗∗
∗∗
∗∗
∗∗
11
5
9
Means followed by the same letter in the same column are not significantly different at the 5%
probability level by Tukey’s test.
∗, ∗∗
Significant at the 5 and 1% probability level, respectively.
However, at high N level, grain yield varied from 37.7 g per pot to 69.6 g
per pot. The increase in grain yield with the application of 400 mg N kg−1
of soil was 354% as compared to treatment without N fertilization across 19
genotypes. This means that upland rice genotypes tested in this study were
not able to extract sufficient N from low N level that can bring significant
grain yield differentiation. Hence, all the genotypes tested were N inefficient
at low N level. Fageria et al., (1995a; 1995b) reported significant yield differences among upland rice genotypes under low, medium and high fertility
levels in Brazilian Oxisol under field conditions.
N and genotypes treatments and their interactions (Table 3) significantly
affected panicle number per pot, panicle length, and grain harvest index.
Panicle number varied from 3.7 to 10.3 per pot at low N level with an average
value across genotypes of 7.9 per pot. Similarly, panicle number at higher
N level varied from 14.3 to 28 per pot with an average value of 21.4 per
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
1702
N. K. Fageria et al.
pot. Overall, the increase in panicle number per pot was 171% with the
application of N as compared to treatment without N fertilization. Number
of panicles (X) were having significant (P < 0.01) quadratic relationship
with grain yield (Y = −32.9349 + 7.2697X – 0.1467X2, R2 = 0.81∗∗ ). This
means 81% variability in grain yield was accounted due to panicle number.
Panicle number per unit area is considered as one of the important yield
components in increasing upland rice yield (Fageria et al., 1997; Fageria,
2007, 2009). Panicle length varied from 13.9 to 20.8 cm with average value
of 17.7 cm at low N level. Similarly, at high N level panicle length varied
from 18.4 to 26.6 cm with an average value of 22 cm. Overall, the increase
in panicle length at the higher N level was 24% compared with low N level.
Panicle length (X) had a significant (P < 0.01) quadratic relationship with
grain yield (Y = −110.5796 + 9.9081X – 0.1299X2, 0.43∗∗ ). Yoshida (1981)
reported that the number of spikelets or grain per unit area of rice crop
is positively correlated with amount of N absorbed by the end of spikelet
initiation stage or by flowering. Figure 1 shows the relationship between
plant height, shoot dry weight, panicle number, panicle length and grain
yield.
Grain harvest index varied from 0.36 to 0.52 with an average value of 0.43
at low N level and 0.33 to 0.57 with an average value of 0.50 at high N level (Table 3). Overall, there was a 16% increase in grain harvest index with the application of N as compared with control treatment. Grain harvest index is an
important parameter in determining distribution of photosynthetic product
between shoot and grain and consequently grain yield (Ishii, 1995). The review on the breeding progress for high yields indicates that the improvement
of the harvest index has made substantial contributions to achieving high
yields of rice (Chandler, 1969). Peng et al. (2000) determined the trend in
the yield of rice cultivars/lines developed since 1966 at the International Rice
Research Institute (IRRI), Philippines. These authors concluded that the increasing trend in yield of cultivars released before 1980 was mainly due to the
improvement in grain harvest index (GHI), while an increase in total biomass
was associated with yield trends for cultivars/lines developed after 1980. They
also suggested that further increases in rice yield potential will likely occur
through increasing biomass production rather than increasing GHI.
Kiniry et al. (2001) reported that the GHI varied greatly among cultivars,
locations, seasons, and ecosystems, ranging from 0.35 to 0.62, indicating the
importance of this variable for yield stimulation. Kiniry et al. (2001) also
reported that yield variability among rice cultivars is highly dependent on
grain harvest index. The importance of this variable is also confirmed when
we compared grain yield of highest yielding genotype CNAs 8993 which
had a grain harvest index of 0.57 and lowest yielding genotypes CNA 8170
had a harvest index of 0.33 at higher N level. This genotype (CNA 8170)
had the highest number of panicles among all the genotypes tested. This
means that there should be a balance among the entire yield attributing
1703
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Nitrogen vs Upland Rice Genotypes
FIGURE 1 Relationship between plant height, shoot dry weight, panicle number, panicle length and
grain yield of upland rice.
parameters of plant to obtain higher yield. Peng et al. (2000) reported that
high yielding rice cultivars/lines developed at the IRRI from 1985 to 1995
were intermediate in canopy height, tillering capacity, and leaf area index.
Cultivars in this group were able to maintain relatively high grain harvest
index.
Nitrogen concentration in shoot and N uptake in shoot (N concentration in shoot X shoot dry weight) was significantly influenced by N and
genotype treatments (Table 4). However, N concentration in grain was not
influenced either by N or by genotype treatments. But N uptake in grain
was only influenced by N treatment. Nitrogen concentration in shoot varied
from 4.8 to 8.9 g kg−1 with an average value of 6.4 g kg−1 at low N level.
1704
N. K. Fageria et al.
TABLE 4 Concentration and uptake of nitrogen in the shoot and grain of nineteen upland rice
genotypes at two nitrogen levels
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
N conc. in shoot
(g kg−1)
N uptake in shoot
(mg pot−1)
Genotype
Low N
level
High N
level
Low N
level
High N
Level
N conc. in
grain (g kg−1)
across two
N levels
CRO 97505
CNAs 8993
CNAs 8812
CNAs 8938
CNAs 8960
CNAs 8989
CNAs 8824
CNAs 8957
CRO 97422
CNAs 8817
CNAs 8934
CNAs 9852
CNAs 8950
CNA 8540
CNA 8711
CNA 8170
BRS Primavera
BRS Canastra
BRS Carisma
Average
6.3abc
6.3abc
6.9abc
5.7bc
6.0bc
6.3abc
4.9c
5.4bc
6.7abc
5.6bc
8.9a
7.2abc
7.6ab
4.8c
6.6abc
5.7bc
6.0bc
7.6ab
6.7abc
6.4
8.7ab
7.0ab
9.2ab
11.3ab
10.4ab
12.8a
9.8ab
10.8ab
9.8ab
10.5ab
10.5ab
10.5ab
9.1ab
9.8ab
10.0ab
8.7ab
7.2ab
5.7b
8.3ab
9.5
100.0abc
93.5abc
117.7abc
79.2bc
89.2bc
91.2bc
68.2c
100.5abc
109.1abc
84.9bc
142.3a
100.7abc
115.4abc
72.1c
111.8abc
99.0abc
87.1bc
123.3abc
94.4abc
98.9
473.1a
363.4a
574.5a
588.1a
478.3a
635.1a
456.9a
481.1a
491.1a
581.9a
620.2a
473.9a
434.6a
578.0a
541.8a
654.3a
352.2a
382.1a
552.4a
511.2
10.9a
11.3a
13.3a
11.9a
12.4a
11.7a
12.6a
12.8a
12.5a
12.1a
11.5a
11.3a
12.6a
12.3a
12.5a
11.1a
12.5a
12.1a
12.8a
12.1
409.0a
513.5a
510.3a
386.3a
429.7a
470.9a
417.4a
465.5a
387.3a
444.0a
370.5a
385.3a
422.0a
405.3a
414.9a
298.1a
484.8a
381.5a
439.4a
422.9
NS
NS
NS
14
NS
NS
25
F-Test
N Level (N)
Genotype (G)
N×G
CV (%)
∗
∗∗
∗∗
∗∗
∗∗
∗∗
16
21
N uptake in
grain (mg pot−1)
across two
N levels
∗∗
Means followed by the same letter in the same column are not significantly different at the 5%
probability level by Tukey’s test.
∗ , ∗ ∗ , NS
Significant at the 5 and 1% probability level and nonsignificant, respectively.
All the genotypes showed visual N deficiency symptoms at low N level. This
means average N concentration of 6.4 g kg−1 in shoot can be considered as
deficient level in upland rice. At higher N level, N concentration in shoot
varied from 5.7 to 12.8 g kg−1 with an average value of 9.5 g kg−1. Visual N
deficiency symptoms were not observed in any genotypes at higher N level.
Hence, average value of 9.5 g N kg−1 can be taken as an adequate concentration in the shoot of upland rice at harvest. Fageria (1998) reported
adequate concentrations for maximum yield about 8.7 g kg−1 in the shoot
of upland rice under field condition at harvest. The slightly higher value of
N concentration in the shoot of present study may be due to use of different
genotypes. Nitrogen concentration in the shoot had a significant quadratic
association with grain yield (Y = −38.3532 + 126.1739X – 41.0378X2,
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Nitrogen vs Upland Rice Genotypes
1705
R2 = 0.34∗∗ ). This means improving N concentration in upland rice genotypes can improve grain yield.
In the grain, N concentration was 11.3 g kg−1 at low N level and 12.8
g kg−1 at high N level across the 19 upland rice genotypes (data not presented). These concentrations can be considered deficient and sufficient for
upland rice, respectively. Cassman et al. (2002) reported that N concentration of about 12 g kg−1 is desired in the grain of rice for optimal cooking
and eating quality. Nitrogen concentration was higher in grain compared to
shoot. Nitrogen concentration in the grain of rice is always higher than the
stover (Kiniry et al., 2001). However, there was less variation in N concentration in grain among genotypes compared to N concentration in the shoot.
Nitrogen concentration in grain had a highly significant positive association
with grain yield (r = 0.31∗∗ ). Hence, increasing N concentration in grain
can increase upland rice grain yield. Nitrogen uptake in shoot varied from
72.1 to 142.3 mg pot−1 at low N level with an average value of 98.9 mg pot−1.
At higher N level, N uptake in the shoot varied from 352.2 to 654.3 mg pot−1
with an average value of 511.2 mg pot−1. Uptake of N in grain varied from
298.1 to 513.5 mg pot−1 across two N levels. Uptake of N in the shoot and
grain followed dry matter yield of these two plant parameters. Uptake of N
in shoot (r = 0.85∗∗ ) as well as in grain (0.93∗∗ ) had a highly significant
associated with grain yield. This indicates that increasing N accumulation in
shoot as well as grain can improve upland rice yield. However, influence of
N accumulation in grain is higher compared to shoot.
Nitrogen use efficiencies defined in Table 1 were calculated for genotypes (Table 5). Agronomical efficiency (AE) and utilization (EU) were
significantly different among genotypes. Physiological efficiency (PE), agrophysiological efficiency (APE) and apparent recovery efficiency (ARE) were
also varied among genotypes, however difference was statistically not significant. The AE varied from 12.8 to 26.7 mg grain produced per mg N applied
with an average value of 21.4 mg grain produced per mg of N applied. Average value across the genotypes of PE was 86.6 mg dry matter production
(grain plus straw) per mg N accumulated in grain and straw. Average value
of APE was 45.2 mg grain produced per mg N accumulated in grain and
straw. Fageria and Barbosa Filho (2001) determined APE in eight lowland
rice genotypes and reported average value of 45.5 mg grain produced per mg
N accumulated in the grain and straw. Hence, APE of lowland and upland
rice is comparable. The average ARE was 49.2%. The ARE in lowland rice is
reported to be in the range of 31 to 40% in major rice growing regions of
the world (Cassman et al. 2002). Fageria and Baligar (2001) reported that
average ARE in lowland rice in Brazilian Inceptisol was 39%. This means
ARE in upland rice is higher compared to lowland rice. The highest grain
yield producing genotype CNAs 8993 had the highest AE, whereas the lowest
yielding genotype CNA 8170 had the lowest AE. The genotype CNAs 8992
also had the highest PE and APE and reasonably good values of ARE and EU.
1706
N. K. Fageria et al.
TABLE 5 Nitrogen use efficiency of nineteen upland rice genotypes
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Genotype
CRO 97505
CNAs 8993
CNAs 8812
CNAs 8938
CNAs 8960
CNAs 8989
CNAs 8824
CNAs 8957
CRO 97422
CNAs 8817
CNAs 8934
CNAs 9852
CNAs 8950
CNA 8540
CNA 8711
CNA 8170
BRS Primavera
BRS Canastra
BRS Carisma
Average
F-Test
Genotype
CV (%)
Agronomical
efficiency
(mg mg−1)
Physiological
efficiency
(mg mg−1)
Agrophysiological
efficiency
(mg mg−1)
Apparent
recovery
efficiency (%)
Utilization
efficiency
(mg mg−1)
23.5ab
26.7a
24.2ab
21.8ab
21.7ab
24.8aab
17.0ab
22.4ab
19.8ab
17.9ab
20.3ab
22.9ab
21.0ab
22.2ab
19.4ab
12.8b
22.6ab
21.7ab
24.2ab
21.4
100.2a
101.4a
80.5a
87.8a
79.3a
72.3a
73.7a
72.3a
85.7a
72.7a
83.3a
81.7a
88.5a
89.7a
79.2a
92.8a
89.8a
125.7a
89.6a
86.6
55.4a
59.0a
41.3a
46.3a
46.1a
42.3a
36.7a
46.3a
47.1a
33.6a
40.5a
48.9a
49.9a
45.2a
40.0a
28.0a
51.2a
57.7a
42.8a
45.2
43.5a
47.8a
58.2a
48.7a
48.3a
59.3a
47.0a
50.7a
45.0a
53.2a
50.2a
47.7a
44.7a
53.5a
48.1a
45.4a
45.8a
40.6a
56.2a
49.2
42.5ab
45.5ab
47.0ab
40.9ab
37.3ab
42.4ab
33.5b
35.2ab
36.6ab
38.3ab
41.8ab
38.4ab
37.6ab
44.1ab
38.0ab
42.0ab
39.9ab
47.4ab
50.5a
41.0
NS
23
NS
26
NS
24
12
∗
17
∗∗
Means followed by the same letter in the same column are not significantly different at the 5%
probability level by Tukey’s test.
∗ , ∗ ∗ , NS
Significant at the 5 and 1% probability level and nonsignificant, respectively.
The AE had a highly significant (P < 0.01) relationship with grain yield
(r = 0.77∗∗ ). The PE was not related significantly with grain yield, however APE had a highly significant (P < 0.01) relationship with grain yield
(r = 0.37∗∗ ). The ARE (r = 0.31∗ ) and UE (r = 0.40∗∗ ) were also associated
significantly with grain yield. Figure 2 shows relationship between grain yield
and different U-use efficiencies for upland rice genotypes.
Selected soil chemical properties of 10 genotypes selected on the basis
of highest, medium and lowest grain yield production were determined at
harvest. Soil pH and extractable soil P were significantly influenced by N
level as well as genotype treatments (Table 6). However, exchangeable Mg
in the soil was only affected significantly by genotype treatment. In addition,
exchangeable Ca, K and H + Al, cation exchange capacity (CEC) and base
saturation were significantly influenced by N level (Table 7). Soil pH varied
from 6.3 to 6.6 with an average value of 6.4 among genotypes. The variation
in pH among genotypes may be related to unbalanced uptake of cations
and anions. When more cations are absorbed, pH of the growth medium
decreases due to release of H+ ions by the roots. While anions absorption
1707
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Nitrogen vs Upland Rice Genotypes
FIGURE 2 Relationship between different N use efficiencies and grain yield.
is higher than cations, OH− or HCO−
3 ions are released and pH increases
(Moore, 1974).
Variation in soil Mg among genotypes was 11.6 to 13.6 with an average
value of 12.5. Similarly, variation in extractable soil P was 21.9 to 40.2 mg
dm−3 with an average value of 31.7 mg dm−3. Fageria and Morais (1987) and
Fageria and Baligar (1997a) have reported variation in upland rice genotypes
for extraction of Mg and P in Oxisols.
Application of 400 mg N kg−1 of soil increases soil pH from 6.2 to 6.6
compared to control treatment or without N application treatment across 10
genotypes (Table 7). The increase in pH may be related to use of urea as N
1708
N. K. Fageria et al.
TABLE 6 Soil pH, extractable soil Mg and P determined after harvest of ten upland rice genotypes
Genotype
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
CRO 97505
CNAs 8993
CNAs 8812
CNAs 8989
CNAs 8824
CRO 97422
CNA 8170
BRS Primavera
BRS Canastra
BRS Carisma
Average
F-Test
N Level (N)
Genotype (G)
N×G
CV(%)
pH in water
Mg (mmol dm−3)
P (mg dm−3)
6.3c
6.4abc
6.4abc
6.4abc
6.6a
6.4abc
6.5ab
6.3c
6.5ab
6.5ab
6.4
11.5c
12.7abc
12.3abc
11.6ac
12.1abc
11.7ac
13.5ab
12.2abc
13.6ab
14.1b
12.5
38.9ad
25.6bc
31.9abcd
34.2abcd
39.9a
40.2a
25.2bc
31.0abcd
28.6bcd
21.9c
31.7
∗∗
NS
∗∗
∗
∗∗
NS
12
NS
26
∗
NS
3
Means followed by the same letter in the same column are not significantly different at the 5%
probability level by Tukey’s test.
∗ , ∗ ∗ , NS
Significant at the 5 and 1% probability level and nonsignificant, respectively.
fertilizer. Urea hydrolysis generates alkalinity by the reaction: CO(NH2 )2 +
+
2H2 O ⇔ HCO−
3 + NH4 + NH3 (Gaudin and Dupuy, 1999). The formation
−
of HCO3 and NH3 ions is responsible for increasing soil pH. Soil pH had
a highly significant positive association with grain yield across 10 genotypes
(r = 0.64∗∗ ). This means increasing soil pH increases grain yield of upland
rice genotypes. Although, upland rice is tolerant to soil acidity (Fageria,
2001b), a positive quadratic relationship was reported with soil pH and grain
yield of upland rice genotypes in Brazilian Oxisol (Fageria, 2000). Fageria
(2001c) reported adequate value of 12 mmol Mg2+ dm−3 in Brazilian Oxisol
for upland rice. Similarly, Fageria (1990) determined 30 to 40 mg P dm−3
TABLE 7 Selected soil chemical properties as influenced by two applied N levels
across 10 upland rice genotypes determined after harvest
Soil chemical
property
pH in water
Ca (mmol dm−3)
P (mg dm−3)
K (mg dm−3)
H + Al (mmol dm−3)
CEC (mmol dm−3)
Base saturation (%)
Low N
(0 mg kg−1)
High N
(400 mg kg−1)
6.2b
25.8b
35.1a
201.7a
31.7a
75.1a
61.2a
6.6a
29.2a
28.4b
37.5b
27.2b
69.9b
57.8b
Means followed by the same letter under two N levels are not significantly different
at the 5% probability level by Tukey’s test.
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Nitrogen vs Upland Rice Genotypes
1709
for 100% relative grain yield of three upland rice cultivars under greenhouse
conditions. This means exchangeable Mg2+ and available P were sufficient
in the soil for upland rice genotypes. This hypothesis was also confirmed by
nonsignificant correlation of exchangeable Mg2+ (r = 0.01) with grain yield
and highly significant negative correlation (r = −0.93∗∗ ) with available P
and grain yield.
Exchangeable soil Ca2+ significantly increased with the application of N,
whereas available P, K, H + Al, CEC and base saturation decreased with increase N level in the soil (Table 7). The increase in exchangeable Ca2+ and
decrease in available P with increasing N level may be related to increase
in soil pH. Fageria (1984) reported that available P in Oxisol of cerrado
region of Brazil decreased when soil pH was raised more than 6.2 due to
formation of insoluble calcium phosphate. Exchangeable soil Ca2+ had a
significant positive correlation (r = 0.45∗ ) with grain yield. Fageria (1984)
reported that upland rice genotypes grain yield increased with the application of calcium carbonate but increase varied with genotype to genotype.
The decrease in soil K at higher level of N was very drastic compared to K
level of control treatment. Potassium had a significant negative correlation
(−0.93∗∗ ) with grain yield. This means that it was not deficient in the soil.
The decrease in K is related to high uptake rate of K by rice genotypes with
the increase of yield at high N level. Potassium uptake was highest by upland
rice compared to all other essential macronutrients (Fageria, 2001a). The
decrease in H + Al, CEC and base saturation at high N level is associated with
decrease in H + Al and K with increasing pH. The H + Al (r = −0.46∗∗ ) and
CEC (r = −0.44∗∗ ) were having significant negative correlation with grain
yield.
CONCLUSIONS
Upland rice genotypes responded differently to applied N at higher N
level. However, at lower N level, grain yield and most of the yield components did not differ significantly. This suggests that in selection upland rice
genotypes use of adequate N rate is essential. Plant height, shoot dry matter,
panicle number, panicle length, grain harvest index, N concentration in
shoot and grain, and N uptake in shoot and grain were having significant
positive relationship with grain yield. These plant parameters were also significantly affected by N as well as genotypes treatments. Hence, it is possible
to manipulate these plant parameters in favor of higher grain yield by using
adequate N rate or planting N efficient genotypes. Based on results of this
study it can be concluded that upland rice average yield can substantially
be increased with planting N efficient genotypes along with improved crop
management practices.
1710
N. K. Fageria et al.
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
REFERENCES
Bremner, J. M., and C. S. Mulvaney. 1982. Nitrogen. In: Methods of Soil analysis, Part 2, 2nd ed., ed. A. L.
Page, pp. 595–624. Madison, WI: American Society of Agronomy.
Cassman, K. G., A. Dobermann, and D. T. Walters. 2002. Agroecosystems, nitrogen-use efficiency, and
nitrogen management. Ambio 31: 132–140.
Chandler, R. F., Jr. 1969. Plant morphology and stand geometry in relation to nitrogen. In: Physiological
Aspects of Crop Yield, eds. J. D. Eastin, F. A. Haskins, C. Y. Sullivan, and C. H. M. Van Bavel, pp.
265–285. Madison, WI: ASA, CSCA.
EMBRAPA. 1997. Manual for Methods of Soil Analysis; 2nd ed. Rio de Janeiro, Brazil: National Research
Center for Soils.
Fageria, N. K. 1984. Response of rice cultivars to liming in cerrado soil. Pesquisa Agropecuaria Brasileira 19:
883–889.
Fageria, N. K. 1990. Phosphorus calibration analysis for rice crop in greenhouse. Pesquisa Agropecuaria
Brasileira 25: 579–586.
Fageria, N. K. 1998. Evaluation nutritional status of rice. In: Technology for Upland Rice, eds. F. Breseghello,
and L. F. Stone, pp. 59–66. Santo Antonio de Goiás, Brazil: National Rice & Bean Research Center
of EMBRAPA.
Fageria, N. K. 2000. Upland rice response to soil acidity in cerrado soil. Pesquisa Agropecuaria Brasileira 35:
2303–2307.
Fageria, N. K. 2001a. Nutrient management for improving upland rice productivity and sustainability.
Communications in Soil Science and Plant Analysis 32: 2603–2629.
Fageria, N. K. 2001b. Effect of liming on upland rice, common bean, corn, and soybean production in
cerrado soil. Pesquisa Agropecuaria Brasileira 36: 1419–1424.
Fageria, N. K. 2001c. Response of upland rice, dry bean, corn and soybean to base saturation in cerrado
soil. Revista Brasileira de Engenharia Agricola e Ambiental 5: 416–424.
Fageria, N. K. 2007. Yield physiology of rice. Journal of Plant Nutrition 30: 843–879.
Fageria, N. K. 2009. The Use of Nutrients in Crop Plants. New York: CRC Press.
Fageria, N. K., and V. C. Baligar. 1997a. Upland rice evaluation for phosphorus use efficiency. Journal of
Plant Nutrition 20: 499–509.
Fageria, N. K., and V. C. Baligar. 1997b. Response of common bean, upland rice, corn,
wheat, and soybean to soil fertility of an Oxisol. Journal of Plant Nutrition 20: 1279–
1289.
Fageria, N. K., and V. C. Baligar. 2001. Lowland rice response to nitrogen fertilization. Communications
in Soil Science and Plant Analysis 32: 1405–1429.
Fageria, N. K., and V. C. Baligar. 2005. Enhancing nitrogen use efficiency in crop plants. Advances in
Agronomy 88: 97–185.
Fageria, N. K., V. C. Baligar, and C. A. Jones. 1997. Growth and Mineral Nutrition of Field Crops. New York:
Marcel Dekker.
Fageria, N. K., and M. P. Barbosa Filho. 2001. Nitrogen use efficiency in lowland rice genotypes. Communications in Soil Science and Plant Analysis 32: 2079–2089.
Fageria, N. K., and O. P. Morais. 1987. Evaluation of rice cultivars for utilization of calcium and magnesium
in the cerrado soil. Pesquisa Agroecuaria Brasileira 22: 667–672.
Fageria, N. K., E. P. Santana, E. M. Castro, and O. P. Morais. 1995a. Differential response of upland rice
genotypes to soil fertility. Revist Brasileira de Ciência do Sol 19: 261–267.
Fageria, N. K., E. P. Santana, and O. P. Morais. 1995b. Response of favourable upland rice genotypes to
soil fertility. Pesquisa Agropecuaria Brasileira 30: 1155–1161.
Gaudin, R., and J. Dupuy. 1999. Ammonical nutrition of transplanted rice fertilized with large urea
granules. Agronomy Journal 91: 33–36.
IRRI (International Rice Research Institute). 1984. Terminology for Rice Growing Environments. Los Banos,
Philippines: International Rice Research Institute.
Ishii, R. 1995. Roles of photosynthesis and respiration in the yield-determining process. In: Science of
the Rice Plant: Physiology, Vol. 2, eds. T. Matsuo, K. Kumazawa, R. Ishii, K. Ishihara, and H. Hirata,
pp. 691–703. Tokyo: Food and Agricultural Policy Research Center.
Kiniry, J. R., G. McCauley, Y. Xie, and J. G. Arnold. 2001. Rice parameters describing crop performance
of four U. S. cultivars. Agronomy Journal 93: 1354–1361.
Nitrogen vs Upland Rice Genotypes
1711
Downloaded By: [Fageria, N.] At: 15:28 27 July 2010
Kiuchi, T., T. Yoshida, and M. Kono. 1966. Relationships between growth luxunance and grain yield in
rice plants. Journal of Agricultural Science 21: 551–554.
Moore, D. P. 1974. Physiological effects of pH on roots. In: The Plant Root and Its Environment, ed. E. W.
Carson, pp. 135–151. Charlottesville, VA: University Press of Virginia.
Peng, S., R. C. Laza, R. M. Visperas, A. L. Sanico, K. G. Cassman, and G. S. Khush. 2000. Grain yield of
rice cultivars and lines developed in the Philippines since 1966. Crop Science 40: 307–314.
Yoshida, S. 1981. Fundamentals of Rice Crop Science. Los Bãnos, Philippines: International Rice Research
Institute.